Claude Code for Beginners: Skills, MCP, Hooks & Real Builds

A step-by-step walkthrough of installing Claude Code across all three interfaces — the desktop app, VS Code, and the terminal — so you're set up no matter which one you chose in Section 1.

A step-by-step walkthrough of installing Claude Code across all three interfaces — the desktop app, VS Code, and the terminal — so you're set up no matter which one you chose in Section 1.

A step-by-step walkthrough of installing Claude Code across all three interfaces — the desktop app, VS Code, and the terminal — so you're set up no matter which one you chose in Section 1.

The everyday permission modes that control how much Claude Code can do on its own, how to switch between them, why "Plan first, then execute" is a habit worth building, and how to redirect or stop it mid-task.

Hands-on practice switching between permission modes in a live session — moving into Plan Mode, approving and switching to Accept Edits, redirecting Claude Code mid-task, and pointing it directly at a file.

How to run more than one Claude Code session at once for real throughput without conflicts, plus how Claude Code can preview and check its own work in the browser as it builds.

Hands-on practice running two Claude Code sessions in parallel on two different tasks, using the browser preview to check work as it happens, then bringing both sessions' results together.

A first look at four tools you'll go deeper on later — scheduled tasks, MCP, skills, and plugins — introduced with plain-English analogies so you know what each one is for before you need it.

Hands-on first contact with all four extension tools — setting up a scheduled task, connecting a real MCP tool, running an existing skill, and installing a plugin from the marketplace.

The most common beginner mistakes from this section (accepting diffs without reading them, vague prompts, sticking to one mode, never using a second session), a quick-reference cheat sheet, and a slightly bigger build to put it all together.

A quick-reference cheat sheet covering everything from this section — permission modes, steering, context, parallel sessions, the browser preview, and automation — followed by a mini-project that pulls it all together into one small build.

What makes a prompt vague versus specific, the three things every good prompt should include, why showing beats just telling, and how breaking work into small steps changes your results.

Hands-on practice turning a vague request into a specific, well-formed prompt — adding a clear goal, constraints, and done-criteria — then a mini-project rebuilding something using nothing but well-written prompts.

What "context" actually means, why more context isn't automatically better, what the context window is and why it matters, and three commands that help you manage it.

Hands-on practice managing context in a real project — checking usage with /context, compacting after more work, and starting fresh with /clear.

How to stop repeating yourself — what CLAUDE MD is, the difference between personal and project memory, and how the /init command sets one up for you in minutes.

Hands-on practice setting up CLAUDE MD from scratch — generating a first draft, customizing it, proving it actually sticks across a session, then writing your own from a blank file.

The recurring mistakes to avoid from this section, a cheat sheet you'll actually use, and a single build that exercises everything — prompting, context, and CLAUDE MD — together.

A checklist covering prompting, context and the context window, and CLAUDE MD basics, followed by a mini-project that puts a well-briefed Claude Code to work end to end.

What slash commands are, the built-ins worth knowing, and how to go beyond them by writing your own custom command from scratch.

Hands-on practice with slash commands — trying the built-ins worth knowing, then writing a custom command from scratch and using it in a small project.

The difference between a skill and a slash command, the anatomy of a skill, where it lives, and when to reach for one instead of the other.

Hands-on practice building a skill from its three core parts, then using it in a small project of your own.

Why and when to delegate a task to a subagent instead of doing it yourself, what comes back when you do, and how to brief a subagent clearly — like briefing a colleague.

Hands-on practice writing a clear brief and delegating a real task to a subagent, then applying it in a small project.

A recap of slash commands, skills, and subagents side by side, common mistakes with each, a cheat sheet, and one build that uses all three together.

What's actually happening when you connect a tool like Gmail, Slack, or Notion to Claude Code, where connectors come from, and how trust and scope work when you sign in.

Hands-on practice connecting a real MCP tool to Claude Code and using it for an actual task, with sample prompts by connector type and a small project to try it end to end.

What's inside a plugin, where plugins come from, how to manage the ones you've installed, and how to choose the right one for what you're trying to do.

Hands-on practice evaluating a plugin before installing it, managing what's already installed, and applying one in a small project.

A genuinely new idea for most beginners: what a hook is, what can trigger one, a concrete example, and a word of caution on how much automation to hand off.

Hands-on practice setting up a real hook from scratch and putting it to work in a small project.

Setting up one-time and recurring tasks, how to write instructions for something that'll run without you watching, and how to manage several scheduled tasks at once.

Hands-on practice setting up a real scheduled task, scheduling one through a connector, and managing your scheduled tasks going forward, with a small project to apply it.

The four building blocks from this section side by side, common mistakes, a cheat sheet, and one automation that pulls together several tools at once.

A cheat sheet covering connectors, plugins, hooks, and scheduled tasks side by side, followed by a mini-project that chains several of them into a small automation pipeline.

A repeatable build process — base prompt, Plan Mode, CLAUDE MD, build, test, iterate — walked through on real featured builds (including a hand-gesture window controller and a 3D solar system explorer), followed by a full catalog of 25+ projects across web apps, Python/computer-vision tools, 3D apps and games, AI applications, and automation utilities, each with a ready-to-use starter prompt. Includes a downloadable bonus: a library of 31 additional project ideas at the same level of ambition.

A single quick-reference walkthrough of the VS Code interface, permission modes and sessions, context, CLAUDE MD, and commands, and everything else — connectors, plugins, skills, subagents, and scheduled tasks — plus 10 additional projects built specifically for this interface.

A go-to reference for the command line, starting from absolute terminal basics through installing and running Claude Code, sessions and modes, context, slash commands, MCP, plugins, skills/subagents, and scheduled tasks — plus 17 additional projects, including computer-vision, 3D, and game builds.

Turning a photo into an AI-generated character image in a riggable A-pose, then bringing it into Meshy to generate, texture, rig, and animate a real 3D model — idle, walk, and run — ready to drop into your own 3D world.

Everything that turns your character into a finished, live portfolio: a browser-based, GTA-style 3D city built with ThreeJS, with third-person controls, AI traffic, drivable cars, and a HUD with a minimap; a coin-based mission system — floating monuments tracked on the map — that reveals a real panel about you, ready to swap in your actual resume; and deploying it all to GitHub Pages, including a security check for exposed secrets before it goes public. Includes downloadable resources: the bonus 31-project idea library, and a short closing message on what to build next.

In this lecture , we will create a Exploration based Resume inside a fantasy world

A bonus library of 31 more project ideas at the same level of ambition as the capstone — immersive 3D worlds, browser games, ambitious Python builds, AI-powered applications, and data visualizations — each with a ready-to-use starter prompt to drop straight into Claude Code.